In this Stanford‑produced video, preventive cardiologist Dr. Fatima Rodriguez frames the growing burden of cardiovascular disease and its inequitable impact through two patient stories: Maria, a 45‑year‑old mother who suffered a heart attack despite no known risk factors, and Doug, a 66‑year‑old who had severe coronary calcification identified years after a routine chest CT but was never notified. Dr. Rodriguez explains that traditional coronary artery calcium (CAC) scans—gated chest CTs—are effective for risk stratification but are limited by insurance coverage and cost, creating access disparities. She then describes a Stanford research program that developed an artificial‑intelligence algorithm to automatically quantify CAC from the millions of non‑gated chest CTs already performed for other reasons (e.g., COVID‑19, lung cancer screening). By scanning historical records, the team identified patients without known heart disease, generated personalized images highlighting abnormal calcification, and sent low‑touch notifications to primary‑care physicians with statin recommendations, followed by patient outreach. In a controlled trial, statin initiation rose from ~7% in the control group to >50% in the notified group, with additional improvements in blood pressure control, diabetes prevention, and a shift from sick‑visit to well‑visit patterns. Doug’s delayed notification experience spurred his involvement as a patient advocate, emphasizing the need for timely communication. Dr. Rodriguez expands the vision to using AI on other imaging modalities (chest X‑ray, abdominal CT) and data sources (ECG, electronic records) and even large language models to reduce clinician burden and detect early disease across organ systems. She concludes with optimism that interdisciplinary AI tools can reverse the upward trend in cardiovascular mortality and promote health equity.
Chapters
Meet Maria. Maria is a 45-year-old busy mother of 3 who finally decides to take control of her own health. She decides to join a local gym— good idea— and start exercising regularly. But about 3 days into her new exercise program, Maria develops burning in her chest while on the treadmill. She thinks it must just be heartburn and continues walking on the treadmill. But the pain intensifies, and Maria collapses to the ground. The staff at the gym recognize something is not right, and they call 911. She's rushed to Stanford Health Care, where she's diagnosed with a heart attack, and she is rushed to the cardiac cath lab, where her occluded coronary artery is opened with a stent. Now, this is a life-saving procedure, and Maria has a long road to recovery. But like Maria, many patients wonder, is there a way I could have known? That I was at risk of heart attack, could this have been prevented? Heart disease is the number one killer of men and women worldwide. This is the big problem that we're trying to solve. Over several decades, we had made strong dents to the heart disease mortality through advances in treatment and prevention. But recently, there has been an ominous upward trajectory as shown here. This is due to an increased burden of risk factors in our population. Equally important is something that we're very passionate about is the inequity. These risk factors are not distributed equally across populations, and we know that certain patient groups are at higher risk of not just risk factors, but also succumbing from these risk factors. A Black individual is 32% more likely to die from heart disease than a white individual. This is a staggering statistic, and we know that it starts early with diagnosis and starting treatment. I'm a preventive cardiologist. That's my disclosure. I spend a lot of my time trying to work with patients to try to personalize their risk and help them make decisions about how to reduce their risk of having a heart attack or stroke. We often rely on traditional risk factors that you're all familiar with, like having high blood pressure or high cholesterol. But sometimes, like in Maria's case, there are no risk factors. In fact, we know that about a third of people who present with a heart attack have never had any symptoms of heart disease. So in my clinic, I often rely on a very valuable tool called a coronary artery calcium scan. This is a specialized type of chest CT scan that is gated to the beating heart. We get very beautiful images of the arteries of the heart, and then we can quantify the amount of calcified plaque in these arteries. As that number gets higher, so does your risk of having a heart attack, and that can help us decide what kind of therapies we want to start to prevent that. Yet we know that those CT scans are not available to everyone because they're not typically covered by health insurance. And there's a modest out-of-pocket cost associated with them. That creates an issue with equity and access. On the other hand, we know that nearly 20 million chest CT scans are done in the United States for reasons other than to assess cardiovascular risk each year. These include things like COVID— remember COVID? COVID pneumonia, or trying to understand lung cancer screening. So could we use this information that's already sitting there in radiology repositories to help us guide preventive therapies. And that's exactly what we did at Stanford. With a team of cardiologists, radiologists, computer scientists, and engineers, we developed an artificial intelligence algorithm that could very quickly and efficiently quantify the amount of coronary calcium in these non-specialized CT scans that were already sitting there. Valuable information that can help us make decisions. But I think what we did next was the real innovation. What we did is that we scanned years of medical records using scans done at Stanford Health Care, identified patients who did not carry a history of heart disease, and then used the algorithm to quantify the amount of plaque in the heart. We created a personalized image such as this one that showed the patient, you have plaque in your arteries, But if that was subtle, we put a big red circle on it. This is not normal. We then notified the primary care physician of the presence of this abnormal finding and made some recommendations about starting preventive medications like statins. Statins are very important medications, my favorite medication as a preventive cardiologist, to lower your cholesterol and your risk of heart disease. If the primary care physician did not object, we then notified the patient 2 weeks later. And what did we find? We found that in those individuals who we did not notify, the control arm of the experiment, around 7% of those individuals started statins. Many of them had higher cardiovascular risk. But in individuals that we notified through this low-touch, high-tech, and personalized intervention, over 50% started statins. That is a really large effect size. And in addition to starting statins, which was the primary outcome of this trial, we noticed that patients were doing other things to improve their heart disease health. They were controlling their blood pressure better, preventing diabetes, and we saw a transition in the type of visits that these patients had from sick visits to well visits, and that's really the goal. We had many patients and their clinicians who sent us messages saying this was a very powerful motivator to start having discussions about reducing heart disease risk. But let me tell you about someone we didn't reach. Meet Doug. Doug is a 66-year-old semi-retired plant breeder from Fresno, California, who 4.5 years ago had a non-gated chest CT to follow up on a history of tonsil cancer. Now, the chest CT results were excellent. Good news, no recurrence of cancer. Bad news, Doug had severe calcification in the arteries of his heart. Even worse news, Doug was not notified for 4.5 years. Now, this is what Doug has to say. Doug felt like the system had failed him. And Doug reached out to our team, and now he is an active member of our research team, meets with us weekly. We even have Doug writing papers. Doug is just the best, and he is confident that technology exists to assist everyone in the system to deliver more accurate and timely care. And he's helping us design our interventions with the patient perspective in mind. This is really only the beginning of how artificial intelligence and technology can be used in clinical medicine to bridge gaps in access, health equity, early diagnosis before disease manifests as severe. We've all heard of large language models like ChatGPT, but can we use those kind of models to help us dig through those medical records? I just heard today about an intervention at Stanford. Could we use this to respond to patient messages in a way to reduce our clinician burden? Or how about something as simple as a chest X-ray, commonly performed? Researchers at Stanford have shown that it can provide valuable information about the presence of diabetes that has not yet been diagnosed. Or a tool that I use all the time as a cardiologist, an electrocardiogram, a bunch of squiggly lines that can actually tell us about cardiac function, structure, and can diagnose inherited cardiac disease before it presents. Our group is also interested in moving beyond just the chest CT and calcification to abdominal imaging that is very frequently performed. And can we use information like tissue composition to diagnose not just heart disease, but liver disease, kidney disease, osteoporosis? The possibilities are really endless. I'll end by saying this is where we are. But I am optimistic that through tools like artificial intelligence, a shared vision, and the unique collaboration of different disciplines, which is really a strength at Stanford, we can reverse this trajectory. And I really look forward to a future where everyone has the opportunity to lead a healthy life free of cardiovascular disease. Thank you.
The preceding program is copyrighted by the Board of Trustees of the Leland Stanford Junior University. Please visit us at med.stanford.edu.
